Upload
nera-logo-lendo
View
17
Download
1
Tags:
Embed Size (px)
DESCRIPTION
Wernicke’s aphasia occurs after a stroke to classical language comprehension regions in the left temporoparietal cortex.Consequently, auditory–verbal comprehension is significantly impaired in Wernicke’s aphasia but the capacity to comprehendvisually presented materials (written words and pictures) is partially spared. This study used functional magnetic resonanceimaging to investigate the neural basis of written word and picture semantic processing in Wernicke’s aphasia, with the wideraim of examining how the semantic system is altered after damage to the classical comprehension regions. Twelve participantswith chronic Wernicke’s aphasia and 12 control participants performed semantic animate–inanimate judgements and a visualheight judgement baseline task. Whole brain and region of interest analysis in Wernicke’s aphasia and control participantsfound that semantic judgements were underpinned by activation in the ventral and anterior temporal lobes bilaterally. TheWernicke’s aphasia group displayed an ‘over-activation’ in comparison with control participants, indicating that anterior temporallobe regions become increasingly influential following reduction in posterior semantic resources. Semantic processing ofwritten words in Wernicke’s aphasia was additionally supported by recruitment of the right anterior superior temporal lobe, aregion previously associated with recovery from auditory-verbal comprehension impairments. Overall, the results provide supportfor models in which the anterior temporal lobes are crucial for multimodal semantic processing and that these regions maybe accessed without support from classic posterior comprehension regions.
Citation preview
BRAINA JOURNAL OF NEUROLOGY
The anterior temporal lobes support residualcomprehension in Wernickes aphasiaHolly Robson,1,2 Roland Zahn,1,3 James L. Keidel,4 Richard J. Binney,1 Karen Sage1,5 andMatthew A. Lambon Ralph1
1 Neuroscience and Aphasia Research Unit, School Psychological Sciences, University of Manchester, UK
2 School of Psychology and Clinical Language Sciences, University of Reading, UK
3 Department of Psychological Medicine, Institute of Psychiatry, Kings College London, UK
4 School of Psychology, Bangor University, UK
5 Faculty of Health and Social Care, University of the West of England, Bristol, UK
Wernickes aphasia occurs after a stroke to classical language comprehension regions in the left temporoparietal cortex.
Consequently, auditoryverbal comprehension is significantly impaired in Wernickes aphasia but the capacity to comprehend
visually presented materials (written words and pictures) is partially spared. This study used functional magnetic resonance
imaging to investigate the neural basis of written word and picture semantic processing in Wernickes aphasia, with the wider
aim of examining how the semantic system is altered after damage to the classical comprehension regions. Twelve participants
with chronic Wernickes aphasia and 12 control participants performed semantic animateinanimate judgements and a visual
height judgement baseline task. Whole brain and region of interest analysis in Wernickes aphasia and control participants
found that semantic judgements were underpinned by activation in the ventral and anterior temporal lobes bilaterally. The
Wernickes aphasia group displayed an over-activation in comparison with control participants, indicating that anterior tem-
poral lobe regions become increasingly influential following reduction in posterior semantic resources. Semantic processing of
written words in Wernickes aphasia was additionally supported by recruitment of the right anterior superior temporal lobe, a
region previously associated with recovery from auditory-verbal comprehension impairments. Overall, the results provide sup-
port for models in which the anterior temporal lobes are crucial for multimodal semantic processing and that these regions may
be accessed without support from classic posterior comprehension regions.
Keywords: Wernickes aphasia; semantic processing; language comprehension; anterior temporal lobe; Wernickes area
Abbreviation: PALPA = psycholinguistic assessment of language processing in aphasia
IntroductionWernickes aphasia is the classical aphasic syndrome associated
with impaired language comprehension. Wernickes aphasia results
from lesions to the left posterior temporoparietal cortex (Bogen
and Bogen, 1976; Dronkers et al., 1995), thereby affecting core
elements of the phonological and semantic systems (Robson et al.,
2012a, 2013) that interact during language comprehension.
The comprehension impairment in Wernickes aphasia is modu-
lated by the degree of phonological analysis required (Robson
et al., 2012b). Spoken language comprehension, which requires
a high degree of phonological analysis for word recognition, is
severely impaired in Wernickes aphasia. Written word compre-
hension, which is mediated by both phonological and visual pro-
cesses, is significantly less impaired in Wernickes aphasia in
comparison with spoken word comprehension. Comprehension
doi:10.1093/brain/awt373 Brain 2014: 137; 931943 | 931
Received April 12, 2013. Revised November 16, 2013. Accepted November 25, 2013 The Author (2014). Published by Oxford University Press on behalf of the Guarantors of Brain.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0/), which permits unrestricted reuse,
distribution, and reproduction in any medium, provided the original work is properly cited.
by guest on June 5, 2015D
ownloaded from
of pictorial materials, which primarily requires visual analysis pro-
cesses, is comparatively preserved (although not necessarily intact)
in Wernickes aphasia (Robson et al., 2012b). Given that semantic
cognition is distributed over a number of perisylvian and extra-
sylvian regions (Jefferies and Lambon Ralph, 2006), it is of clinical
and neurobiological importance to determine the functional
anatomy that supports residual semantic processing of visually-
presented items (pictures and written words) in a group of homo-
genous, classical Wernickes aphasia participants.
Recent functional imaging studies involving participants with
variable levels of recovered aphasia have demonstrated the im-
portance of anterior temporal lobe regions for language compre-
hension, in particular the anterior fusiform gyrus, superior anterior
temporal lobe and temporal pole. Sharp et al. (2004) used an
auditory semantic association task in a functional PET investigation
to assess individuals with lesions to the left superior temporal lobe
who had recovered single word comprehension. Increased activa-
tion was found in the right anterior fusiform gyrus and bilateral
temporal poles. Functional MRI studies using passive narrative lis-
tening have also emphasized the role of the superior anterior tem-
poral lobes. Activation in the right superior anterior temporal lobe
during narrative listening was found to correlate with measures of
sentence comprehension in individuals with a history of aphasia
(Crinion and Price, 2005) and the degree of interhemispheric func-
tional connectivity between the superior anterior temporal lobes
was found to correlate with single word and sentence comprehen-
sion measures in participants with chronic aphasia (Warren et al.,
2009).
The evidence that the anterior temporal lobes support compre-
hension in aphasia recovery is consistent with other lines of re-
search that highlight the anterior temporal lobes as crucial
components of the semantic network; specifically, regions that
support abstraction of transmodal representational semantic
knowledge (Patterson et al., 2007; Lambon Ralph et al., 2010).
The strongest evidence for anterior temporal lobe involvement in
semantic representation has emerged from neuropsychological in-
vestigations of semantic dementia patients, who have progressive
atrophy focussed on the anterior temporal lobes bilaterally and a
resultant progressive degradation of semantic representations
(Mummery et al., 2000; Patterson et al., 2007). Additionally, re-
petitive transcranial magnetic stimulation to the right or left an-
terior temporal lobe in neurologically-normal young controls slows
reaction times for semantic but not numerical judgements
(Lambon Ralph et al., 2009; Pobric et al., 2010a, b) and neuroi-
maging studies (which avoid or correct for signal dropout) have
shown responses in the anterior temporal lobes during multimodal
semantic processing tasks (Vandenberghe et al., 1996; Marinkovic
et al., 2003; Liu et al., 2009; Binney et al., 2010; Visser et al.,
2010a).
These lines of research suggest that anterior temporal regions
support core conceptual knowledge and that these areas may
become increasingly influential when posterior areas of the lan-
guage network are damaged. Such notions contrast with trad-
itional neurobiological models of language (Geschwind, 1965;
Geschwin, 1972). Based on deficits associated with Wernickes
aphasia, traditional models have emphasized the role of the pos-
terior temporoparietal region as the central access point to, or
locus for the representation of semantic knowledge. Current
neurobiological models of language implicate a wider, distributed
lateral and medial temporalparietalfrontal network, within which
the temporoparietal region remains a core element (Binder et al.,
2009).
Although we have emphasized here the importance of anterior
temporal lobe regions in semantic representation, it should be
noted that the hub-and-spoke model of semantic representation
(Rogers et al., 2004; Patterson et al., 2007; Lambon Ralph et al.,
2010) builds on previous notions that concepts derive from anter-
ior temporal lobe coordinated activation of information encoded in
a set of distributed modality-specific association areas (the
spokes). Meynert and Wernickes view of conceptualization sug-
gested that only the distributed modality-specific regions were ne-
cessary. Direct descendants of these ideas are found in the
modern literature and captured in the hypothesis of embodied
cognition (Barsalou et al., 2003), which can vary in form from
weak to strong formulations (Meteyard et al., 2012). Again, the
key idea in these theories is that concepts reflect the mass action
of multiple information sources that are experienced and encoded
in each modality, separately. Considerable evidence for this ap-
proach has come from functional neuroimaging and neuropsycho-
logical studies (Pulvermuller, 2005; Martin, 2007; Kemmerer et al.,
2012). The hub-and-spoke hypothesis suggests that coherent con-
cepts require both transmodal anterior temporal lobe representa-
tions plus these distributed modality-specific sources of
information (for discussion of these issues, see: Patterson et al.,
2007; Lambon Ralph, 2013). The combined roles of transmodal
anterior temporal lobe and modality-specific regions in semantic
processing have been confirmed by utilizing transcranial magnetic
stimulation to investigate and compare different neural regions
within the same neurologically-intact participants (Pobric et al.,
2010a).
In contrast with the majority of neuroimaging studies of apha-
sia, this study focused on a group of individuals with classical
chronic Wernickes aphasia. We used distortion-corrected func-
tional MRI to investigate semantic processing with the aims of
revealing important insights into the clinical manifestation of
Wernickes aphasia and the neural basis of semantic cognition.
Specifically, we explored: (i) the neural regions underlying seman-
tic processing in Wernickes aphasia; and (ii) how the semantic
system adapts to the removal of core posterior components.
Participants with chronic Wernickes aphasia and age-matched
controls made semantic judgements about single items which
engage minimal or moderate phonological processing (pictures
and written words, respectively, in comparison with spoken
word processing). Based on previous findings in aphasia and the
theory that the anterior temporal lobes support representational
semantics, it was hypothesized that participants with Wernickes
aphasia should demonstrate significant activation of the anterior
temporal lobe when semantically processing visually-presented
materials (picture or written words). Because lesions in
Wernickes aphasia affect phonological processing regions of the
left superior temporal lobe, it was hypothesized that written words
processing would show a greater degree of reorganization in com-
parison with picture processing and engage superior temporal re-
gions of the right hemisphere. We tested these hypotheses in
932 | Brain 2014: 137; 931943 H. Robson et al.
by guest on June 5, 2015D
ownloaded from
12 individuals with classical Wernickes aphasia and 12 control
participants. To ensure complete and reliable coverage of all an-
terior temporal regions, we used distortion-corrected functional
MRI which minimizes signal loss and distortion over these areas
(Embleton et al., 2010).
Materials and methods
Participant diagnosis and lesionsTwelve participants with chronic Wernickes aphasia [two female,
mean age 70.1, standard deviation (SD) 8.7] and 12 age- and educa-
tion-matched control participants (one female, mean age 71, SD 6.9)
were recruited and provided written informed consent as approved by
the Multicentre Research Ethics Committee NHS ethics committee. All
participants were right-handed in that they wrote with their right
hand. Control participants were screened to ensure they had no pre-
vious or current neurological, language or cognitive deficit and all were
native speakers of English.
All patients presented with classical symptoms of Wernickes aphasia
after a single left hemisphere stroke; namely impaired single word
comprehension, single word repetition and fluent, sentence-like
speech punctuated with phonological or neologistic errors. Diagnosis
was confirmed using the Boston Diagnostic Aphasia Examination
(Goodglass et al., 2001). Table 1 displays biographical and behavioural
diagnostic data. Additional background language data were collected
for: single-word spoken versus written comprehension; written word
versus picture semantic association judgements; and single word read-
ing aloud. The results are displayed in Table 2. Single word compre-
hension was assessed using the word-to-picture matching test from
the 64-item Cambridge Semantic Battery (Bozeat et al., 2000). In this
test a spoken or written word is presented and the participant is asked
to select the matching item from a set of 10 semantically-related pic-
tures. The spoken and written versions use the same items and differ
only on modality of presentation. Semantic association judgements
were assessed using the Pyramids and Palm Trees test (Howard and
Patterson, 1992). In this test the participant must judge which of two
semantically related items is associated with a probe item. Two ver-
sions of the test were administered, in which the same triads are pre-
sented either as written words or pictures. Single-word reading aloud
was taken from the Psycholinguistic Assessment of Language
Processing in Aphasia (PALPA: Kay et al., 1992). Consistent with pre-
vious reports (Robson et al., 2012a), test accuracy varied with the
degree of phonological processing required, so that semantic associ-
ation judgements were significantly more accurate for pictures than
written words [t(11) = 2.8, P = 0.017] and single word comprehension
was significantly more accurate for written than spoken words
[t(10) = 5.4, P5 0.001]. Table 2 summarizes the results from theseassessments for each participant. All participants displayed impaired
spoken single word comprehension and 7/11 participants displayed
impaired written word comprehension. Seven participants were im-
paired at picture semantic association and eight participants were im-
paired at written word semantic association. The participants who
were unimpaired at semantic association displayed the least severe
comprehension impairment overall.
Structural T1-weighted magnetic resonance images were acquired
before functional MRI scanning on a 3 T Philips Achieva scanner
with an eight-element SENSE head coil and a sense factor of 2.5.
An inversion recovery sequence produced a 256 256 matrix of128 transverse slices with 1 mm3 voxels. Lesions were extracted Tab
le1
Dem
ogra
phic
and
dia
gnost
icas
sess
men
tfo
rW
ernic
kes
aphas
iaan
dco
ntr
ol
par
tici
pan
ts
WA
Age
Sex
Tim
epost
-onse
t(m
onth
s)
Aet
iolo
gy
Lesi
on
BD
AE
com
pre
hen
sion
BD
AE
fluen
cyB
DA
Ere
pet
itio
nC
ontr
ol
Age
Sex
AC
E-R
Tota
l%
Tota
lm
ax32
%Se
nte
nce
%W
ord
%M
MSE
Tota
lM
ax30
max
100
DR
76
M7
InpST
L,TPJ,
IFL
26
47
51
51
BR
76
M26
78*
DM
C67
M10
Hae
mST
G,
TPJ,
MTL,
IFL,
IPL
34
47
51
51
BH
67
M30
98
DL
73
M9
InST
L,M
TL,
ATL,
IFL
35
63
51
51
DC
74
M27
85
LS66
M10
InpST
L,pM
TL,
IPL
510.5
70
25
25
DW
72
M30
92
LB80
F84
InST
L,M
TL,
TPJ,
IPL
59
68
515
EC78
F28
87
CB
59
M14
InST
L,IP
L,IF
L,A
TL
10
838
15
10
TT
61
M26
91
RD
87
M17
InpST
L,IP
L,TPJ
10
14
80
510
NJ
78
M30
88
MC
73
F13
InpST
L,TPJ,
IPL
10
14
83
10
10
HE
76
M30
98
EL61
M15
InpST
L,M
TL,
TPJ
14
17
75
10
15
ML
66
M28
90
NM
59
M11
InST
L,M
TL,
TPJ,
ATL
17
15.5
100
10
10
AM
58
M30
96
CH
77
M17
InST
L,M
TL,
TPJ
40
25
90
45
5G
P78
M26
93
CW
70
M36
InST
L,M
TL,
TPJ,
IPL
40
26
100
40
15
KW
69
M30
99
Wer
nic
kes
aphas
ia(W
A)
par
tici
pan
tsw
ere
scre
ened
and
dia
gnose
dusi
ng
the
BD
AE
(Bost
on
Dia
gnost
icA
phas
iaEx
amin
atio
n).
Contr
olp
artici
pan
tsw
ere
scre
ened
usi
ng
AC
E-R
=A
dden
bro
oks
Cognitiv
eEx
amin
atio
n
Rev
ised
,fr
om
whic
ha
Min
i-M
enta
lSt
ate
Exam
inat
ion
(MM
SE)
score
(Fols
tein
etal.,
1975)
can
be
der
ived
.W
ernic
kes
aphas
iapar
tici
pan
tsar
eord
ered
by
seve
rity
of
auditory
com
pre
hen
sion
dis
ord
er,
most
(DR
)to
leas
t(C
W).
Aet
iolo
gy:
In=
infa
rct;
Hae
m=
hae
morr
hag
e.Le
sion
indic
ates
core
regio
ns
of
left
hem
ispher
eco
rtic
alan
dsu
bco
rtic
aldam
age
inth
eW
ernic
kes
aphas
iapar
tici
pan
ts:p=
post
erio
r;ST
L=
super
ior
tem
pora
llobe;
TPJ
=te
mporo
par
ieta
ljunct
ion;
IFL
=in
ferior
fronta
llo
be;
MTL
=m
iddle
tem
pora
llo
be;
ATL
=an
terior
tem
pora
llo
be;
IPL
=in
ferior
par
ieta
llo
be.
*C
ontr
olpar
tici
pan
tsw
ith
low
leve
llit
erac
ysc
ore
.
The anterior temporal lobes support residual comprehension in Wernickes aphasia Brain 2014: 137; 931943 | 933
by guest on June 5, 2015D
ownloaded from
using the automated lesion identification algorithm (Seghier et al.,
2008) and overlaid to produce a lesion overlap map (Fig. 1).
Maximal lesion overlap occurred in the white matter underlying the
posterior superior temporal lobe, consistent with the classical descrip-
tion of Wernickes aphasia (Bogen and Bogen, 1976). However, in no
participant was lesion location isolated to the superior temporal lobe;
in all participants the lesion extended into the inferior parietal lobe/
temporoparietal junction, eight of the participants has significant
middle temporal lobe extension and four of the (most severely af-
fected) participants had further extension into the inferior frontal
lobe (Table 1 and Fig. 1).
Functional magnetic resonance imagingtasksSuccessful functional imaging in chronic aphasic participants requires
the selection of tasks which are achievable given their level of impair-
ment (Price and Friston, 1999). Accordingly, this study adapted a se-
mantic anterior temporal lobe-activating paradigm used in a previous
study with young controls (Visser and Lambon Ralph, 2011) to be
suitable for severely impaired participants. The functional MRI tasks
consisted of two animateinanimate judgements tasks using the same
items, one in a pictorial and one in a written word modality, and one
Table 2 Background language testing in Wernickes aphasia group
PicturePyramidsand PalmTrees
WrittenPyramidsand PalmTrees
Writtenword-to-picturematch
Spokenword-to-picturematch
PALPAsinglewordreading
PT Max 52 52 64 64 80Cut-off 49 49 63 63 73
DR 47 33 33 9 0
DMC 42 39 28 16 1
DL 46 32 54 8 14
LS 46 34 N/A 32 0
LB 48 42 58 26 14
CB 42 43 50 30 0
RD 50 52 64 47 4
MC 50 47 60 55 21
EL 48 36 59 30 27
NM 52 52 62 53 28
CH 51 50 63 53 19
CW 51 52 64 51 44
Mean 47.8 42.7 54.1 34.2 14.3
SD 3.3 7.8 12.5 17.4 14.1
Table displays background behavioural semantic and comprehension assessments. Italics indicate outside normal limits. PT = participant. ThePyramids and Palm Trees test (Howard and Patterson, 1992) assesses semantic association, word-picture-matching (Bozeat et al., 2000) assessessingle word comprehension. Reading score from subtest 31 of PALPA (Kay et al., 1992). N/A = not available.
Figure 1 Lesion overlap map for the 12 participants with Wernickes aphasia. The lesion distribution mirrors previous studies ofWernickes aphasia, with lesions centred on posterior perisylvian cortical and subcortical regions. Colour bar indicates the number of
participants with a lesion at each voxel (min = 3; max = 12).
934 | Brain 2014: 137; 931943 H. Robson et al.
by guest on June 5, 2015D
ownloaded from
visual judgement baseline task in which participants judged whether
scrambled pictures were high or low on the screen. Each task (picture,
word or pattern presentation) was presented in a separate run; they
were not mixed because of potential difficulties in task switching that
occur in stroke aphasia (Jefferies and Lambon Ralph, 2006). Each run
was 9.6 min long and consisted of 16 blocks of stimuli (4 s/stimulus
plus 0.5 s prestimulus fixation, total 18 s) interspersed with 16 blocks of
rest (18 s). A total of 64 stimuli were presented per run (32 animate,
32 inanimate). Participants responded using a two button response
(animate versus inanimate, high versus low). The button box was
placed near the participants left hand and participants were able to
practice responding with this button box before going into the scan-
ner. All pre-scan training was undertaken using a left-handed re-
sponse. Once a response was made the stimulus was removed from
the screen and replaced with a fixation point. The stimuli for the ani-
mate-inanimate judgements were the same in both modalities and
consisted of eight semantic categories: animate (domestic animals,
exotic animals, birds and insects) and inanimate (toys, small household
items, clothes and instruments). All stimuli had a spoken word fre-
quency of 520 occurrences per million (Celex database: Baayenet al., 1995) and an imageability of 4500 (maximum possible image-ability 700: Gilhooly and Logie, 1980; Bird et al., 2001), obtained
through the N-Watch database (Davis, 2005). Pictures were black
and white and primarily sourced from the Snodgrass and
Vanderwart set (1980). Scrambled picture stimuli were created by
dividing each stimulus into 80 pieces and randomly redistributing
them using the Java Runtime Environment (www.SunMicrosystems.
com). Pictures, words and scrambled pictures were centred on either
the top or bottom third of the screen. Stimulus presentation was ran-
domized within task and the order of the tasks was randomized be-
tween participants. Tasks were delivered using the E-Prime 1.2
software programme (Psychological Software Tools). All control par-
ticipants received the same training for the three tasks, a total of
15 min. The Wernickes aphasia participants received individualizedpackages of training based on their level of impairment. Training was
stopped when the Wernickes aphasia participants could achieve above
80% accuracy outside the scanner environment on all tasks without
external facilitation by the trainer. Stimuli used in the training did not
overlap with the experiment stimuli. Inside and outside the scanner,
instructions were presented in a pictorial format because of signifi-
cantly impaired comprehension in the Wernickes aphasia group.
MR-compatible eyeglasses were provided for those who required
them.
Functional magnetic resonance imagingacquisitionThis study used distortion-corrected, spin echo EPI functional MRI.
Spin echo sequences produce spatial distortion (but not drop-out) in
the anterior temporal lobes and orbitofrontal cortices, which can be
corrected using a post-acquisition distortion-correction algorithm
(Embleton et al., 2010). The spin echo EPI sequence included 41
slices covering the whole brain with echo time = 70 ms, repetition
time = 4150 ms, flip angle = 90, 96 96 matrix, reconstructed reso-lution 2.5 2.5 mm, and slice thickness 3.0 mm. A total of 139 timepoints were collected each run in a single phase encoding direction
(KL). To compensate for the distortion, spatial re-mapping correction
was applied following spatial realignment; this required the acquisition
of an additional scan with the participant at rest, consisting of 20
volumes of interleaved dual direction phase encoding (10 left-to-
right phase encoding, KL, and 10 right-to-left phase encoding, KR).
This additional scan was acquired between the first and second experi-
mental run. This method has been successfully applied with
non-neurologically impaired participants to reveal semantically-related
activations in ventrolateral anterior temporal lobe regions (Binney
et al., 2010; Visser et al., 2010a) and is described in full elsewhere
(Embleton et al., 2010).
Functional magnetic resonance imaginganalysis
Preprocessing
Statistical analysis was carried out using the SPM8 software.
Preprocessing and general linear model specification were optimized
for brains with lesions. The registered, distortion-corrected images
were further preprocessed by co-registration to anatomical images
and normalization to MNI space using the unified segmentationnor-
malization procedure (Ashburner and Friston, 2005). This normaliza-
tion procedure was found to produce optimum results for functional
MRI analysis of brains with lesions when medium regularization was
applied (Crinion et al., 2007). Following normalization, images were
smoothed with an 8 mm full-width at half-maximum Gaussian filter.
General linear model analysis is dependent on accurate convolution
with the haemodynamic response function. Previous investigations
have shown the time-to-peak of the blood oxygen level-dependent
response can be significantly delayed in stroke aphasia (Bonakdarpour
et al., 2007). The haemodynamic response function time-to-peak was
analysed in the Wernickes aphasia group using finite impulse response
functions and showed no significant deviations in the areas of peak
activation. This may be a consequence of peak activation areas being
supplied by the non-infarcted posterior cerebral artery (see Results
section). However, time derivatives were added to the general linear
model to account for small deviations in the haemodynamic response
function time-to-peak which may be common in elderly participants
(DEsposito et al., 2003) but not observable with a finite impulse re-
sponse analysis based on a long repetition time of 4150 ms.
Functional magnetic resonance imagingcontrastsAt the first level, whole brain univariate analyses, thresholded at
P5 0.005 with a minimum four-voxel extent, were performed forthe semantic animate-inanimate task in each modality. Semantic
blocks were contrasted against active baseline blocks and rest blocks.
This dual-baseline was considered necessary to ensure that the results
were a true reflection of task-related semantic processing. The contrast
against baseline blocks accounted for activations related to motor,
executive-decision making and visual processes. Rest was used as a
contrast in order to account for default processing (such as day-
dreaming) which may have been present during the semantic blocks.
The task design resulted in periods of rest/fixation during the semantic
blocks as semantic items were removed after participant response.
Default processes engage a network which overlaps with the neural
regions of interest to the current study, including anterior temporal
regions and further frontal regions of the semantic network.
Therefore, this additional rest contrast ensured activations reflected
task-related semantic processing.
In the second-level analysis, the task activation coefficient maps
described above for pictures and written words were entered into a
2 2 mixed-effects ANOVA, with group (Wernickes aphasia versusControl) as a between-subject factor and task and the group task
The anterior temporal lobes support residual comprehension in Wernickes aphasia Brain 2014: 137; 931943 | 935
by guest on June 5, 2015D
ownloaded from
interaction as within-subject factors. Thus, for the main effect of task
and the Group Task interaction, a subject factor was included inSPMs Flexible Factorial module, whereas this factor was not included
for the test of the main effect of Group. Results are displayed for
clusters significant at P5 0.005 uncorrected, minimum cluster sizefour voxels.
Region of interest analysisFurther analyses were carried out over a priori regions of interest
within the sylvian and extra-sylvian semantic network. All regions of
interest were bilateral in order to investigate potential reorganization
or hemispheric lateralization. Following the procedure described in
Visser and Lambon Ralph (2011), regions of interest were derived
from independent literatures (see Table 3 for region of interest coord-
inates and literature sources). Five bilateral region of interest pairs
were investigated: (i) anterior fusiform gyri; (ii) temporal poles; (iii)
anterior superior temporal gyri/sulci; (iv) ventral occipital-temporal
lobe; and (v) inferior frontal gyri. Fusiform gyri, temporal pole and
superior temporal gyri regions of interest were selected following pre-
vious literature indicating their key involvement in semantic processing
in (recovered) aphasia. Inferior frontal gyri regions of interest were
selected because of the functional integration between inferior frontal
gyri regions and posterior temporal semantic regions damaged in the
Wernickes aphasia group. The ventral occipital-temporal lobe regions
of interest were included because of the consistent neural responses
observed in controls in this portion of the ventral visual stream during
word recognition and visual object processing (Chao et al., 1999;
Twomey et al., 2011). Mean beta weights over the regions of interest
were extracted using the MarsBar toolbox (Brett et al., 2002) and
further analysed in SPSS. For each region of interest, 2 2 ANOVAswere used to investigate main effects of group and condition and
group condition interactions and one-sample t-tests were employedto identify regions of significant activation.
Results
Behavioural taskThe Wernickes aphasia comprehension-impaired group was sig-
nificantly less accurate than the control group on the semantic and
baseline tasks. Table 4 displays d accuracy scores and reaction
time data for the Wernickes aphasia and control groups. For
the picture task, 10 of 12 participants with Wernickes aphasia
performed above chance (DL, LS, LB, CB, RD, MC, EL, NM, CH
and CW: binomial P50.001 for all participants). For the wordcondition, eight of the participants with Wernickes aphasia per-
formed above chance (DL, CB, RD, MV, EL NM, CH, CW:
P50.005). For the control condition, eight of the Wernickesaphasia participants performed significantly above chance (DR,
DL, RD, MC, EL, NM, CH, CWL: P5 0.005).
Functional magnetic resonance imagingThree versions of the functional MRI analysis were run based on
the behavioural results: (i) all responses from all participants; (ii) all
responses from participants performing above chance; and (iii)
only correct responses from all participants. Results from analysis
(i) are presented as all analyses yielded similar results. One signifi-
cant difference occurred in the region of interest analysis between
analysis (i) and analysis (ii), reported below.
Whole brain analysis results
Figure 2 displays whole brain results for the Wernickes aphasia
group and regions significantly more active in the Wernickes
aphasia than control group. Peaks for the within-group simple main
effects and main effect of group are presented in Tables 57.
Contrasts of semantic judgements and dual baseline in the
Wernickes aphasia group produced extensive activation through-
out the temporal lobes (Fig. 2). Picture and written word judge-
ments produced activation in the ventral temporal lobes bilaterally.
Picture judgements produced activation bilaterally throughout the
fusiform extending into the temporal poles; bilaterally in the ven-
tral occipito-temporal region and bilaterally in the anterior superior
and middle temporal gyri. Additional peaks were observed in the
medial frontal cortex bilaterally. Written word judgements acti-
vated the bilateral fusiform gyri, left parahippocampal gyrus; left
middle temporal gyrus extending into the temporal pole; right
middle and superior temporal gyri and superior temporal sulcus
extending into the temporal pole. Further peaks were observed
Table 3 Region of interest analysis coordinates
Region of interest MNI coordinates Radius* mm Coordinate source
x y z
Left anterior fusiform gyrus 38 18 32 5 Sharp et al. (2004)Right anterior fusiform gyrus 38 18 32 5 HomologueLeft temporal pole 42 16 32 5 Sharp et al. (2004)Right temporal pole 40 20 34 5 Sharp et al. (2004)Left anterior superior temporal sulcus 54 6 16 7 Scott et al. (2000)Right anterior superior temporal sulcus 54 6 16 7 HomologueLeft inferior frontal gyrus 51 30 6 7 Visser and Lambon Ralph (2011)Right inferior frontal gyrus 50 30 6 7 Homologue
Left ventral occipital-temporal lobe 38 44 18 7 Visser and Lambon Ralph (2011)Right ventral occipital-temporal lobe 42 44 18 7 Visser and Lambon Ralph (2011)
*After Visser and Lambon Ralph (2011) regions of interest were selected from Sharp et al. (2004) and Scott et al. (2000); the same radii dimensions were used for theseregions of interest as in Visser and Lambon Ralph (2011). The larger radius of 7 mm was used for the regions of interest selected from the Visser and Lambon Ralph (2011)
results.
936 | Brain 2014: 137; 931943 H. Robson et al.
by guest on June 5, 2015D
ownloaded from
in the left inferior angular gyrus, left inferior frontal gyrus and
bilateral medial frontal lobe. In contrast, the control group
showed considerably less activation overall with pictures activating
the left posterior fusiform gyrus and words producing activation in
the posterior fusiform gyri bilaterally and left parahippocampal
gyrus. Small additional peaks were observed in the left middle
and superior frontal gyri and right medial frontal lobe.
Corresponding to this, the main effect of group was reflected in
significantly greater activation in the Wernickes aphasia group
than the control group throughout the ventral and middle tem-
poral lobes, and to a greater extent on the left than the right.
Additionally, the Wernickes aphasia group, produced greater ac-
tivation in the right superior temporal sulcus, extending into the
temporal pole. The control group showed significantly greater ac-
tivation than the Wernickes aphasia group only in a single cluster
in the right posterior cingulate. Main effects of condition were
limited. Pictures produced significantly greater activation in the
left ventral temporal pole whereas written words produced
Table 4 Functional MRI behavioural task analysis
Pictures Words Scrambled pictures
d (SD) RT (SD) d (SD) RT (SD) d (SD) RT (SD)Max 13.9 Max 13.9 Max 13.9
Wernickes aphasia 4.9 (5.5) 1460 (422) 4.3 (5.6) 1583 (456) 6.5 (6.0) 1331 (568)
Control 9.8 (4.8) 929 (187) 11.0 (5.3) 1123 (422) 11.7 (4.2) 749 (205)
t-test t(22) 2.3 4 3 2.6 2.5 3.4P 0.03 0.001 0.006 0.018 0.02 0.003
Table displays means and standard deviations for d scores and reaction times for each functional MRI tasks, along with independent samples t-tests displaying group
differences.RT = reaction time.
Figure 2 Whole-brain results for semantic condition versus dual baseline. (A) significant activation in the Wernickes aphasia group forpicture (red) and word (green) semantic judgement, overlapping regions in yellow. (B) Main effect of group. Regions significantly more
active in Wernickes aphasia than control group are shown in blue. No temporal lobe regions were more active in the control than
Wernickes aphasia group. Displayed activations significant at P5 0.005, uncorrected.
The anterior temporal lobes support residual comprehension in Wernickes aphasia Brain 2014: 137; 931943 | 937
by guest on June 5, 2015D
ownloaded from
significantly greater activation in small clusters in the bilateral an-
terior parahippocampal gyri and right inferior frontal gyrus. There
were no Group Condition interactions.
Region of interest analysis
The region of interest analysis (Fig. 3) showed a very similar pat-
tern of results. One tailed one-sample t-tests found that control
participants significantly activated the left anterior fusiform gyrus
for picture semantic judgements [t(11) = 3.1, P = 0.005] and the
left anterior superior temporal gyri/sulci for written word judge-
ments [t(11) = 1.9, P = 0.037]. The Wernickes aphasia group
showed significant activation for both picture and word conditions
in the left anterior fusiform gyrus [pictures: t(11) = 2.8, P = 0.009,
words: t(11) = 2.0, P = 0.034], the right anterior fusiform gyrus
[pictures: t(11) = 2.9, P = 0.008; words: t(11) = 1.9, P = 0.04],
left temporal pole [pictures: t(11) = 2.0, P = 0.034; words:
t(11) = 1.9, 0.04] and the left ventral occipital-temporal lobe [pic-
tures: t(11) = 3.6, P = 0.004; words: t(11) = 2.8, P = 0.017]. The
Wernickes aphasia group displayed additional activation in the
right anterior superior temporal gyrus/sulcus for word stimuli
[t(11) = 2.27, P = 0.022] and in the right ventral occipital-temporal
lobe for the picture stimuli [t(11) = 3.5, P = 0.005]. Removal of the
participants who performed at chance for the picture or written
word judgements made no change to significance with the excep-
tion that significant activation was additionally found in the right
temporal pole for the picture judgements [t(9) = 2.1, P = 0.03]. A
Table 5 Peak whole-brain coordinates for participants with Wernickes aphasia
Task Region Subregion Brodmann area MNI coordinates Z-score
x y z
Pictures4 dual-baseline Temporal lobe L. post. middle temporal gyrus 19 34 80 20 3.12L. ant. middle temporal gyrus 21 56 10 10 2.76R. ant. middle temporal gyrus 21 46 0 22 4.02R. temporal pole 38 48 12 26 3.86
Frontal lobe L. medial frontal gyrus 11 2 26 14 3.67R. medial frontal gyrus 11 2 36 16 3.84R. inferior frontal gyrus 45 50 28 8 3.13
46 42 40 6 2.96
Occipital lobe R. lingual gyrus 17 10 90 0 3.02Cerebellum L. post. lobe 34 64 18 4.96
R. post. lobe 8 52 6 2.83Words4 dual-baseline Temporal lobe L. mid. middle temporal gyrus 21 60 42 12 5.23
L. post. middle temporal gyrus 39 40 66 26 3.71Frontal lobe L. inferior frontal gyrus 47 34 34 18 3.78
L. superior frontal gyrus 11 20 42 12 2.93L. medial frontal gyrus 11 2 36 16 3.06R. medial frontal gyrus 11 8 34 22 3.3R. inferior frontal gyrus 47 22 16 24 2.96
Parietal lobe L. angular gyrus 39 50 68 34 2.79Occipital lobe L. precuneus 31 22 80 26 2.82
R. lingual gyrus 17 12 88 2 3.06Cerebellum L. ant. lobe 2 54 20 2.99
Table 6 Peak whole-brain coordinates for control participants
Task Region Sub-region Brodmannarea
MNI coordinates Z-score
x y z
Pictures4 dual-baseline Temporal lobe L. post. fusiform gyrus 19 38 74 18 3.62Cerebellum L. post. lobe 42 68 22 3.03
R. post. lobe 8 72 36 2.98Words4 dual-baseline Temporal lobe R. post. fusiform gyrus 19 38 74 18 4.13
R. mid. fusiform gyrus 37 50 48 24 3.28L. post. fusiform gyrus 37 42 58 20 3.27L. mid parahippocampal gyrus 35 24 20 14 2.94
Frontal lobe L. middle frontal gyrus 9 40 12 34 3.88R. medial frontal gyrus 25 4 14 18 3.06L. superior frontal gyrus 10 24 52 30 3
938 | Brain 2014: 137; 931943 H. Robson et al.
by guest on June 5, 2015D
ownloaded from
Table 7 Peak whole brain coordinates for Wernickes aphasia4 control
Region Sub-region Brodmann area MNI coordinates Z-score
x y z
Temporal lobe L. uncus 20 32 14 26 3.69L. ant. middle temporal gyrus 21 48 2 28 3.29L. ant. inferior temporal gyrus 20 52 14 24 3.28L. ant. fusiform gyrus 20 54 6 24 3.05R. ant. middle temporal gyrus 21 46 0 22 3.2R. temporal pole 38 48 12 26 2.99R. ant. superior temporal gyrus 21 58 10 16 2.72R. ant. parahippocampal gyrus 35 24 10 22 2.77R. mid. fusiform gyrus 20 46 26 20 2.71
Frontal lobe L. middle frontal gyrus 11 32 36 16 2.77Cerebellum L. post. lobe 18 66 18 3.75
R. post. lobe 44 62 16 3.72
Figure 3 Region of interest analyses. Graphs display mean beta values for each group in each region of interest. Patients with Wernickesaphasia (WA) demonstrated significant activation for picture and written word semantic decisions in the anterior fusiform gyrus and
ventral occipital-temporal lobe bilaterally and in the left lateral polar region. Written word decision produced additionally significant results
in the right anterior superior temporal gyri/sulci. aFuG = anterior fusiform gyrus; aSTG = anterior superior temporal gyrus; TP = temporal
pole; vOT = ventral occipital-temporal lobe; iFrG = inferior frontal gyrus. *P50.05, one sample t-test.
The anterior temporal lobes support residual comprehension in Wernickes aphasia Brain 2014: 137; 931943 | 939
by guest on June 5, 2015D
ownloaded from
ANOVA (2 2) revealed main effects of group in the left ventraloccipital-temporal lobe [F(1,22) = 10.14, P = 0.004], right ventral
occipital-temporal lobe [F(1,22) = 5.81, P = 0.025] and left tem-
poral pole [F(1,22) = 4.65, P = 0.042] caused by significantly
greater semantic activation in the Wernickes aphasia group than
in the control group. There were no significant effects of condition
or significant interactions.
DiscussionBy using distortion-corrected functional MRI, we were able to es-
tablish that individuals with chronic Wernickes aphasia activated
extensive areas of the temporal lobe bilaterally while making
simple superordinate semantic judgements of written words and
pictures. Written words and pictures elicited similar activation pro-
files, despite greater behavioural impairments in written word
comprehension being a characteristic of Wernickes aphasia.
These results provide further evidence for the crucial role played
by the ventrolateral aspects of the anterior temporal lobes in
multimodal semantic processing.
Anterior temporal lobe function inWernickes aphasiaThe Wernickes aphasia group displayed activation through the
inferior temporal lobes bilaterally while making animate-inanimate
judgements. Further peaks were identified in additional compo-
nents of the left hemisphere semantic network in the left inferior
frontal gyrus and left posterior temporoparietal junction. The visual
ventral stream is thought to be organized as a recurrent hierarch-
ical network that enables the formation of successively complex
representations through combining low level visual inputs (Kravitz
et al., 2013). Anterior in the ventral stream, the anterior fusiform
gyri and temporal poles are highly interconnected to other modal-
ity specific processing regions and medial temporal lobe areas
(Duffau et al., 2008; Binney et al., 2012). This interconnectivity
is thought to allow the abstraction of transmodal semantic repre-
sentations (Patterson et al., 2007; Lambon Ralph et al., 2010).
The fact that we observed activation in the anterior temporal
lobe despite posterior temporal lesions provides further evidence
for the key role played by this region in semantic representation in
patients and controls (Sharp et al., 2004; Binney et al., 2010;
Lambon Ralph et al., 2010; Pobric et al., 2010a; Visser and
Lambon Ralph, 2011; Visser et al., 2012). Crucially, the
Wernickes aphasia group showed significantly greater and more
extensive activation than the control group throughout the ventral
and superior anterior temporal lobe; activation increased in the
anterior temporal lobe and other regions following lesions to pos-
terior temporal regions classically associated with semantic repre-
sentation access. This increased activation was bilateral but to a
greater extent on the left. Therefore, the Wernickes aphasia
group appeared to require greater recruitment of regions asso-
ciated with both representational semantic and visual representa-
tions of concepts to perform the semantic task.
The activation pattern found in the Wernickes aphasia group in
the current study mirrored that in a previous study of young
control participants when performing a more challenging, speeded
version of the same semantic judgement task (Visser and Lambon
Ralph, 2011). However, the activation in the Wernickes aphasia
group extended into the temporal poles, particularly in the left
hemisphere, which was not observed in either the elderly or
young controls. Extension into lateral polar cortex has been
observed, however, for more demanding verbal and non-verbal
semantic association decisions (Visser et al., 2012).
Given that control participants recruit similar areas under more
challenging conditions, the results in the Wernickes aphasia group
could be interpreted primarily as an enhancement or overactivity,
rather than recruitment of new neural components. Two accounts
of enhancement have been proposed: (i) enhancement reflects
increased cognitive demands on the region; or (ii) enhancement
reflects disinhibition at the neuronal level as a result of deafferen-
tation from another neural region (Price and Friston, 1999). Both
of these interpretations can be accommodated in the context of
Wernickes aphasia and models of semantic processing. Lesions in
Wernickes aphasia affect both auditoryphonological and seman-
tic processing regions of the posterior language network. In par-
ticular, lesions frequently extend into the left angular gyrus, mid
and posterior middle temporal gyrus, and underlying white matter
(Fig. 1). These regions have been implicated in semantic process-
ing and may support functions such as retrieving and integrating
conceptual information (Lau et al., 2008; Binder et al., 2009).
Some authors have also suggested that this region is important
for extracting conceptual information in relation to events or other
time-varying stimuli, thus supporting associative semantic repre-
sentations (Schwartz et al., 2011). These temporo-parietal junction
regions are often found to co-activate with inferior frontal areas
during semantic processing (Noonan et al., 2010, 2013) and it is
proposed that they function within a cognitive control network
which, within the semantic domain, may regulate task-directed
use of core semantic representations (Jefferies and Lambon
Ralph, 2006; Corbett et al., 2009). Therefore, lesions to posterior
regions may place greater demands on inferior and anterior as-
pects of the extra-sylvian temporal semantic network through re-
duction in semantic processing recourses and/or may lead to
disinhibition with these areas through reduced regulation from
control processes, leading to overactivation of semantic informa-
tion irrelevant for task completion. Because the additional activa-
tion observed in the inferior temporal lobes in the Wernickes
aphasia group replicates previous findings from more challenging
tasks in control participants, it appears that these are not newly
recruited regions. Rather, these extra-sylvian regions are function-
ing in an inefficient and redundant fashion (Price and Crinion,
2005; Zahn et al., 2006); consequently, maximum processing cap-
acity will be reached more rapidly as task difficulty increases.
Reorganization of semantic processingFor the written word semantic judgements, the Wernickes aphasia
group displayed significant activation along the right anterior su-
perior temporal sulcus and superior temporal gyrus (whole brain
and region of interest analysis). In comparison, the control partici-
pants displayed activation in the homologue region, the left an-
terior superior temporal gyrus/sulcus, for the same task (region of
940 | Brain 2014: 137; 931943 H. Robson et al.
by guest on June 5, 2015D
ownloaded from
interest analysis). The left anterior superior temporal sulcus/super-
ior temporal gyrus is associated with the processing of abstract
intelligible phonological information in both spoken and written
forms (Scott et al., 2006; Spitzyna et al., 2007; Richardson
et al., 2011). Dynamic causal modelling of functional MRI data
indicates that the left anterior superior temporal sulcus receives
input from both higher-order visual regions in the ventral occipital
temporal lobe and from auditory-phonological regions in the pos-
terior superior temporal sulcus (Richardson et al., 2011). This con-
verges with psycholinguistic models which propose that written
word comprehension (in semi-transparent orthographies such as
English) is mediated by a phonological as well as a visual pathway
(Plaut et al., 1996). The importance of the right anterior superior
temporal gyrus/sulcus in language comprehension subsequent to
left posterior temporal lesions has been demonstrated in previous
studies of spoken language comprehension and has been related
to the degree of recovery from comprehension impairment
(Crinion and Price, 2005; Warren et al., 2009). In comparison to
previous studies, the current Wernickes aphasia group displayed
written word comprehension impairments and were significantly
less accurate at the functional MRI task than the control group,
but still activated the right anterior superior temporal gyrus/sulcus.
This may indicate that recovery of function is not only related to
which areas of the brain are newly recruited after lesion, but also
to the deeper functional organization of those regions with respect
to factors including the nature of the stimuli, the precision of
perceptual processing of the stimuli and the connectivity of the
regions.
It was hypothesized that, because of the additional phonological
impairment in Wernickes aphasia, the written word semantic
judgements might display a greater degree of reorganization
than the picture judgements. The results did not support this pre-
diction. The region of interest analysis did not reveal any signifi-
cant effects of condition and no Group Condition interactionwas observed at the whole brain level. There may be multiple
explanations for this. The participants with Wernickes aphasia
might have relied to a greater extent on the visual reading path-
way than the phonological pathway to achieve written word
judgements. Conversely, they might have attempted to recruit
phonological regions to support the picture semantic judgements,
e.g. through sub-vocal naming. Given that written word judge-
ments can be achieved through both visual and phonological
routes, further investigation is required to establish whether se-
mantic network recruitment in Wernickes aphasia can be depend-
ent on input modality or is multimodal as suggested by the current
results.
With the exception of the right anterior superior temporal
gyrus/sulcus, the Wernickes aphasia group did not recruit any
additional regions in comparison with the activation profile
observed in young controls undertaking a more demanding ver-
sion of the same task (Visser and Lambon Ralph, 2011). For ex-
ample, the Wernickes aphasia group did not display significant
widespread upregulation of inferior frontal or temporoparietal
junction components of the left hemisphere semantic network or
their right hemisphere homologues. Overall, there was limited evi-
dence for large-scale reorganization of the semantic network in
the chronically impaired individuals who participated in this study.
However, given the low semantic demands of the animate-inani-
mate judgement task, this interpretation must be viewed with
caution. Further work investigating more semantically complex
judgements is required to investigate organizational changes to
the wider semantic system. However, such an approach is highly
methodologically challenging due to the significant behavioural
impairments associated with more complex processing (see
below).
Methodological considerationsThe anterior inferior temporal lobes have often been omitted from
traditional neurobiological models of language. This exclusion may
have multiple methodological causes (Visser et al., 2010; Lambon
Ralph, 2014). The anterior inferior temporal lobes lie outside the
territory of the left middle cerebral artery (Phan et al., 2005). As
middle cerebral artery territory lesions most frequently lead to
aphasia, it is areas within this territory that have been most con-
sistently implicated in language processing, to the exclusion of
extra-middle cerebral artery regions. Secondly, functional MRI
scanning with gradient echo sequences, the most common neu-
roimaging methodology, suffers from signal drop-out and distor-
tion in the orbito-frontal cortex and anterior inferior portions of
the temporal lobes (Weiskopf et al., 2006). Thirdly, neuroimaging
experiments that do not have full brain coverage can cut-off an-
terior inferior temporal lobe regions, if anterior commissurepos-
terior commissure oriented acquisition windows/PET cameras are
aligned to include the top of the brain (Visser et al., 2010b). This
study demonstrated semantic processing in the anterior temporal
areas with a distortion-corrected spin-echo sequence that compen-
sates for signal dropout in the orbitofrontal cortex and anterior
temporal lobes (Embleton et al., 2010).
Task selection is an important consideration when interpreting
neuroimaging results wtih neurologicaly impaired populations. It
has been emphasized that control and patient groups should be
engaging in the same cognitive operations and, therefore, should
be undertaking the same task (Price and Friston, 1999).
Furthermore, the patient group should be able to perform the
task accurately, as a significant degree of error means that the
elicited activation cannot be associated reliably with the intended
cognitive operations but might also reflect alternative/additional
cognitive processes such as error monitoring (Price and Friston,
1999). This study investigated semantic processing of single visu-
ally presented items, comprehension of which is relatively spared
in Wernickes aphasia compared with comprehension of spoken
items. However, despite using a task with low executive and
semantic demands and extensive pre-scan training, performance
in the Wernickes aphasia group remained worse than that of
controls. The Wernickes aphasia group was impaired relative to
the control group on the baseline as well as the semantic tasks.
Although the baseline task has no semantic/linguistic content, this
pattern of results is not unexpected and may reflect multiple
sources of difficulty. As described above, lesions in Wernickes
aphasia extend into the posterior components of a frontoparietal
executive control network required for efficient task completion,
even tasks with limited executive complexity. Additionally, individ-
uals with stroke aphasia have been found to suffer from
The anterior temporal lobes support residual comprehension in Wernickes aphasia Brain 2014: 137; 931943 | 941
by guest on June 5, 2015D
ownloaded from
domain-general attention impairments (Tseng et al., 1993) which
may have contributed to the behavioural impairment. Given these
task differences between the Wernickes aphasia and the control
group it is important to note that the additional analyses, where
poorly-performing or inaccurate patients responses were removed,
generated the same pattern of results.
Finally, although this study addresses individuals with chronic
Wernickes aphasia, for whom no further rapid behavioural im-
provement has been observed, it is important to note that five
participants with Wernickes aphasia were 51 year post-onset atthe time of scanning. As such, some limited plasticity-related
changes may still occur in these participants and the core findings
from this study should be interpreted in this light. To confirm and
enhance the interpretation of results from this study, further lon-
gitudinal neuroimaging work in aphasia is required.
ConclusionThis study found significant activation in anterior temporal lobe
regions during single-item semantic processing in individuals with
classical Wernickes aphasia and, to a lesser extent, in elderly con-
trol participants. The patients reduction in posterior temporal se-
mantic and phonological processing resources increased reliance
on extra-sylvian temporal regions. To enhance the interpretation
of findings in aphasia neuroimaging, future work should attempt
to investigate patterns of response within regions as well as
changes to large-scale network recruitment.
AcknowledgementsWe would like to thank the participants and their relatives/carers for
their assistance in this study, the speech and language therapists in
the North of England who have referred participants to this project
and the 3T MRI radiography team at Hope Hospital, Salford.
FundingThis work was supported by a Stroke Association Allied Health
Professional Research Bursary (TSAB2008/01) and a Stroke
Association Senior Research Training Fellowship grant (TSA SRTF
2012/02) to H.R. and an MRC programme grant to MALR (MR/
J004146/1).
ReferencesAshburner J, Friston KJ. Unified segmentation. Neuroimage 2005; 26:
83951.Baayen R, Piepenbrock R, Gulikers L. The CELEX lexical database
(CR-ROM). Philadelphia: Linguistics Data Consortium, University of
Pennsylvania; 1995.
Barsalou LW, Simmons WK, Barbey AK, Wilson CD. Grounding concep-
tual knowledge in modality-specific systems. Trends Cogn Sci 2003; 7:8491.
Binder JR, Desai RH, Graves WW, Conant LL. Where is the semantic
system? A critical review and meta-analysis of 120 functional neuroi-
maging studies. Cereb Cortex 2009; 19: 276796.
Binney RJ, Embleton KV, Jefferies E, Parker GJM, Lambon Ralph MA.
The ventral and inferolateral aspects of the anterior temporal lobe are
crucial in semantic memory: evidence from a novel direct comparison
of distortion-corrected fMRI, rTMS, and semantic dementia. Cereb
Cortex 2010; 20: 272838.
Binney RJ, Parker GJM, Lambon Ralph MA. Convergent connectivity and
graded specialization in the rostral human temporal lobe as revealed
by diffusion-weighted imaging probabilistic tractography. J Cogn
Neurosci 2012; 24: 19982014.Bird H, Franklin S, Howard D. Age of acquisition and imageability ratings
for a large set of words, including verbs and function words. Behav
Res Methods Instrum Comput 2001; 33: 739.
Bogen JE, Bogen GM. Wernickes region - Where is it? Ann N Y Acad Sci
1976; 280: 83443.
Bonakdarpour B, Parrish TB, Thompson CK. Hemodynamic response
function in patients with stroke-induced aphasia: implications for
fMRI data analysis. Neuroimage 2007; 36: 32231.Bozeat S, Lambon Ralph MA, Patterson K, Garrard P, Hodges JR. Non-
verbal semantic impairment in semantic dementia. Neuropsychologia
2000; 38: 120715.
Brett M, Anton JL, Valabregue R, Poline JB. Region of interest analysis
using the MarsBar toolbox for SPM 99. Neuroimage 2002; 16: S497.
Chao LL, Haxby JV, Martin A. Attribute-based neural substrates in tem-
poral cortex for perceiving and knowing about objects. Nat Neurosci
1999; 2: 9139.
Corbett F, Jefferies E, Ehsan S, Lambon Ralph MA. Different impairments
of semantic cognition in semantic dementia and semantic aphasia:
evidence from the non-verbal domain. Brain 2009; 132: 2593608.Crinion J, Price CJ. Right anterior superior temporal activation predicts
auditory sentence comprehension following aphasic stroke. Brain
2005; 128: 285871.
Crinion J, Ashbumer J, Leff A, Brett M, Price C, Friston K. Spatial nor-
malization of lesioned brains: performance evaluation and impact on
fMRI analyses. Neuroimage 2007; 37: 86675.DEsposito M, Deouell LY, Gazzaley A. Alterations in the bold FMRI
signal with ageing and disease: a challenge for neuroimaging. Nat
Rev Neurosci 2003; 4: 86372.
Davis CJ. N-Watch: a program for deriving neighborhood size and other
psycholinguistic statistics. Behav Res Methods 2005; 37: 6570.
Dronkers NF, Redfern BB, Ludy CA. Lesion localization in chronic
Wernickes aphasia. Brain Lang 1995; 51: 625.
Duffau H, Thiebaut de Schotten M, Mandonnet E. White matter func-
tional connectivity as an additional landmark for dominant temporal
lobectomy. J Neurol Neurosurg Psychiatry 2008; 79: 4925.
Embleton KV, Haroon HA, Morris DM, Lambon Ralph MA, Parker GJM.
Distortion correction for diffusion-weighted MRI tractography and
fMRI in the temporal lobes. Hum Brain Mapp 2010; 31: 157087.Folstein MF, Folstein SE, McHugh PR. Mini-mental state": a practical
method for grading the cognitive state of patients for the clinician. J
Psychiatr Res 1975; 12: 18998.
Geschwind N. Language and the brain. Sci Am 1972; 226: 7683.
Geschwind N. Disconnexion syndromes in animals and man. Brain 1965;
88: 237.
Gilhooly KJ, Logie RH. Age-of-acquisition, imagery, concreteness, famil-
iarity and ambiguity measures for 1,944 words. Behav Res Methods
Instrum 1980; 12: 395427.
Goodglass H, Kaplan E, Barresi B. Boston diagnostic aphasia examination
(BDAE). 3rd edn. Baltimore, MD: Baltimore Lippincott Williams and
Wilkins; 2001.Howard D, Patterson K. Pyramids and palm trees: a test of semantic
access from pictures and words. Bury Saint Edmunds, UK: Thames
Valley Test Company; 1992.
Jefferies E, Lambon Ralph MA. Semantic impairment in stroke aphasia
versus semantic dementia: a case-series comparison. Brain 2006; 129:
213247.
Kay J, Coltheart M, Lesser R. Psycholinguistic Assessments of Language
Processing in Aphasia (PALPA). Hove, UK: Laurence Erlbaum
Associates; 1992.
942 | Brain 2014: 137; 931943 H. Robson et al.
by guest on June 5, 2015D
ownloaded from
Kemmerer D, Rudrauf D, Manzel K, Tranel D. Behavioral patterns andlesion sites associated with impaired processing of lexical and concep-
tual knowledge of actions. Cortex 2012; 48: 82648.
Kravitz DJ, Saleem KS, Baker CI, Ungerleider LG, Mishkin M. The ventral
visual pathway: an expanded neural framework for the processing ofobject quality. Trends Cogn Sci 2013; 17: 2649.
Lambon Ralph MA. Neurocognitive insights on conceptual knowledge
and its breakdown. Philos Trans R Soc B: Biological Sciences 2014;
369: 20120392.Lambon Ralph MA, Pobric G, Jefferies E. Conceptual knowledge is
underpinned by the temporal pole bilaterally: convergent evidence
from rTMS. Cereb Cortex 2009; 19: 8328.Lambon Ralph MA, Sage K, Jones RW, Mayberry EJ. Coherent concepts
are computed in the anterior temporal lobes. Proc Natl Acad Sci USA
2010; 107: 271722.
Lau EF, Phillips C, Poeppel D. A cortical network for semantics: (de)con-structing the N400. Nat Rev Neurosci 2008; 9: 92033.
Liu H, Agam Y, Madsen JR, Kreiman G. Timing, timing, timing: fast
decoding of object information from intracranial field potentials in
human visual cortex. Neuron 2009; 62: 28190.Marinkovic K, Dhond RP, Dale AM, Glessner M, Carr V, Halgren E.
Spatiotemporal dynamics of modality-specific and supramodal word
processing. Neuron 2003; 38: 48798.
Martin A. The representation of object concepts in the brain. Annu RevPsychol 2007; 58: 2545.
Meteyard L, Rodriguez Cuadrado S, Bahrami B, Vigliocco G. Coming of
age: a review of embodiment and the neuroscience of semantics.Cortex 2012; 48: 788804.
Mummery CJ, Patterson K, Price CJ, Ashburner J, Frackowiak RSJ,
Hodges JR. A voxel-based morphometry study of semantic dementia:
relationship between temporal lobe atrophy and semantic memory.Ann Neurol 2000; 47: 3645.
Noonan KA, Jefferies E, Corbett F, Lambon Ralph MA. Elucidating the
nature of deregulated semantic cognition in semantic aphasia: evi-
dence for the roles of prefrontal and temporoparietal cortices. JCogn Neurosci 2010; 22: 1597613.
Noonan KA, Jefferies E, Visser M, Lambon Ralph MA. Going beyond
inferior prefrontal involvement in semantic control: Evidence for theadditional contribution of dorsal angular gyrus and posterior middle
temporal cortex. J Cogn Neurosci 2013; 25: 182450.
Patterson K, Nestor PJ, Rogers TT. Where do you know what you know?
The representation of semantic knowledge in the human brain. NatRev Neurosci 2007; 8: 976.
Phan TG, Donnan GA, Wright PM, Reutens DC. A digital map of middle
cerebral artery infarcts associated with middle cerebral artery trunk and
branch occlusion. Stroke 2005; 36: 98691.Plaut DC, McClelland JL, Seidenberg MS, Patterson K. Understanding
normal and impaired word reading: Computational principles in
quasi-regular domains. Psychol Rev 1996; 103: 56115.
Pobric G, Jefferies E, Lambon Ralph MA. Category-specific versus cat-egory-general semantic impairment induced by transcranial magnetic
stimulation. Curr Biol 2010a; 20: 96468.
Pobric G, Jefferies E, Lambon Ralph MA. Amodal semantic representa-tions depend on both anterior temporal lobes: evidence from repetitive
transcranial magnetic stimulation. Neuropsychologia 2010b; 48:
133642.
Price CJ, Friston KJ. Scanning patients with tasks they can perform. HumBrain Mapp 1999; 8: 1028.
Price CJ, Crinion J. The latest on functional imaging studies of aphasic
stroke. Curr Opin Neurol 2005; 18: 42934.
Pulvermuller F. Brain mechanisms linking language and action. Nat RevNeurosci 2005; 6: 57682.
Richardson FM, Seghier ML, Leff AP, Thomas MSC, Price CJ. Multiple
routes from occipital to temporal cortices during reading. J Neurosci2011; 31: 823947.
Robson H, Sage K, Lambon Ralph MA. Wernickes aphasia reflects a
combination of acoustic-phonological and semantic control deficits: a
case-series comparison of Wernickes aphasia, semantic dementia andsemantic aphasia. Neuropsychologia 2012a; 50: 26675.
Robson H, Keidel JL, Lambon Ralph MA, Sage K. Revealing and quan-
tifying the impaired phonological analysis underpinning impaired com-
prehension in Wernickes aphasia. Neuropsychologia 2012b; 50:27688.
Robson H, Grube M, Lambon Ralph MA, Griffiths TD, Sage K.
Fundamental deficits of auditory perception in Wernickes aphasia.
Cortex 2013; 49: 180822.Rogers TT, Lambon Ralph MA, Garrard P, Bozeat S, McClelland JL,
Hodges JR, et al. Structure and deterioration of semantic memory: a
neuropsychological and computational investigation. Psychol Rev2004; 111: 20535.
Schwartz MF, Kimberg DY, Walker GM, Brecher A, Faseyitan OK,
Dell GS, et al. Neuroanatomical dissociation for taxonomic and the-
matic knowledge in the human brain. Proc Natl Acad Sci USA 2011;108: 85204.
Scott SK, Blank CC, Rosen S, Wise RJS. Identification of a pathway for
intelligible speech in the left temporal lobe. Brain 2000; 123: 24006.
Scott SK, Rosen S, Lang H, Wise RJS. Neural correlates of intelligibility inspeech investigated with noise vocoded speecha positron emission
tomography study. J Acoust Soc Am 2006; 120: 107583.
Seghier ML, Ramlackhansingh A, Crinion J, Leff AP, Price CJ. Lesion
identification using unified segmentation-normalisation models andfuzzy clustering. Neuroimage 2008; 41: 125366.
Sharp DJ, Scott SK, Wise RJS. Retrieving meaning after temporal lobe
infarction: The role of the basal language area. Ann Neurol 2004; 56:83646.
Snodgrass JG, Vanderward M. A standardized set of 260 pictures: norms
for name agreement, image agreement, familiarity and visual complex-
ity. J Exp Psychol Hum Learn 1980; 6: 174215.Spitzyna GA, Wise RJS, McDonald SA, Plant GT, Kidd D, Crewes H, et al.
Optokinetic therapy improves text reading in patients with hemianopic
alexiaa controlled trial. Neurology 2007; 68: 192230.
Tseng CH, McNeil MR, Milenkovic P. An investigation of attention allo-cation deficits in aphasia. Brain Lang 1993; 45: 27696.
Twomey T, Kawabata Duncan KJ, Price CJ, Devlin JT. Top-down modu-
lation of ventral occipito-temporal responses during visual word rec-ognition. Neuroimage 2011; 55: 124251.
Vandenberghe R, Price C, Wise R, Josephs O, Frackowiak RSJ. Functional
anatomy of a common semantic system for words and pictures.
Nature 1996; 383: 2546.Visser M, Lambon Ralph MA. Differential contributions of bilateral ven-
tral anterior temporal lobe and left anterior superior temporal gyrus to
semantic processes. J Cogn Neurosci 2011; 23: 312131.
Visser M, Embleton KV, Jefferies E, Parker GJ, Lambon Ralph MA. Theinferior, anterior temporal lobes and semantic memory clarified: novel
evidence from distortion-corrected fMRI. Neuropsychologia 2010a; 48:
168996.
Visser M, Jefferies E, Lambon Ralph MA. Semantic processing in theanterior temporal lobes: a meta-analysis of the functional neuroima-
ging literature. J Cogn Neurosci 2010; 22: 108394.
Visser M, Jefferies E, Embleton KV, Lambon Ralph MA. Both the middletemporal gyrus and the ventral anterior temporal area are crucial for
multimodal semantic processing: distortion-corrected fMRI evidence
for a double gradient of information convergence in the temporal
lobes. J Cogn Neurosci 2012; 24: 176678.Warren JE, Crinion JT, Lambon Ralph MA, Wise RJS. Anterior temporal
lobe connectivity correlates with functional outcome after aphasic
stroke. Brain 2009; 132: 342842.
Weiskopf N, Hutton C, Josephs O, Deichmann R. Optimal EPI param-eters for reduction of susceptibility-induced BOLD sensitivity losses: a
whole-brain analysis at 3 T and 1.5 T. Neuroimage 2006; 33:
493504.Zahn R, Schwarz M, Huber W. Functional activation studies of word
processing in the recovery from aphasia. J Physiol Paris 2006; 99:
37085.
The anterior temporal lobes support residual comprehension in Wernickes aphasia Brain 2014: 137; 931943 | 943
by guest on June 5, 2015D
ownloaded from